Haitao Lv
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Haitao Lv has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
Yong Wang
University of Electronic Science and Technology of China (UESTC),School of Resources and EnvironmentChengdu,Sichuan,China,611731
8 shared publicationsYang Shen
3 shared publications- Yong Wang
School of Resources and Environment, University of Electronic Science and Technology of China, Chengdu, China
3 shared publications - Jiang Qian
School of Resources and Environment, University of Electronic Science and Technology of China, Chengdu, China
2 shared publications - Minfeng Xing
School of Resources and Environment, University of Electronic Science and Technology of China, Chengdu, China
2 shared publications Yuanyuan Yang
School of Resources and Environment University of Electronic and Science Technology of China (UESTC) 2006 Xiyuan Avenue, West Hi-tech Zone, Chengdu, Sichuan 611731, China
2 shared publications
Latest Publications
- Open Access
- 27 August 2026
- Published by Elsevier in SSRN Electronic Journal
LB-DLPU: A Recoverability-Certified L-Band InSAR Phase-Unwrapping Benchmark for Reliable Deformation Retrieval
Surface deformation associated with groundwater withdrawal, tectonic strain, and volcanic unrest is routinely measured with interferometric synthetic aperture radar (InSAR). The expanding availability of L-band observations from missions such as NISAR and ALOS-4 increases deformation monitoring capability over vegetated terrain but also raises the importance of robust phase unwrapping (PU). A single unresolved phase cycle introduces approximately 12 cm of spurious line-of-sight displacement at L-band. Learning-based unwrappers are increasingly capable, yet existing public PU benchmarks are generally not calibrated to specific L-band instruments, typically omit the ionospheric phase screen, and do not verify that the reference phase is recoverable from the wrapped observation. We present LB-DLPU, a benchmark designed to evaluate PU as an upstream source of uncertainty in L-band deformation estimation. Every synthetic scene passes a recoverability certificate: a noiseless oracle minimum-cost-flow test that retains only scenes recovered within a stated tolerance after removal of a best-fitting real-valued constant offset. Distinct NISAR and UAVSAR regimes are calibrated from real interferograms in terms of coherence, posting, phase noise, and, for NISAR, the large-scale ionospheric spectrum. The benchmark also provides per-edge five-arc ambiguity labels, difficulty strata, and whole-tile DEM splits that prevent terrain leakage. Simulated residue densities are close to or moderately higher than those measured in the calibration data. Across eleven classical, network-flow, and deep-learning baselines, performance separates strongly by sensor regime and difficulty. A model trained only on LB-DLPU transfers without adaptation to real NISAR and UAVSAR interferograms; on a San Andreas UAVSAR scene, its output agrees with the JPL operational product to within one fringe over 95.9% of the scene. LB-DLPU, its frozen splits, generation code, and evaluation harness are publicly released to support reproducible assessment of L-band deformation-retrieval reliability.
- 27 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessLB-DLPU: A Recoverability-Certified L-Band InSAR Phase-Unwrapping Benchmark for Reliable Deformation Retrieval
Surface deformation associated with groundwater withdrawal, tectonic strain, and volcanic unrest is routinely measured with interferometric synthetic aperture radar (InSAR). The expanding availability of L-band observations from missions such as NISAR and ALOS-4 increases deformation monitoring capability over vegetated terrain but also raises the importance of robust phase unwrapping (PU). A single unresolved phase cycle introduces approximately 12 cm of spurious line-of-sight displacement at L-band. Learning-based unwrappers are increasingly capable, yet existing public PU benchmarks are generally not calibrated to specific L-band instruments, typically omit the ionospheric phase screen, and do not verify that the reference phase is recoverable from the wrapped observation. We present LB-DLPU, a benchmark designed to evaluate PU as an upstream source of uncertainty in L-band deformation estimation. Every synthetic scene passes a recoverability certificate: a noiseless oracle minimum-cost-flow test that retains only scenes recovered within a stated tolerance after removal of a best-fitting real-valued constant offset. Distinct NISAR and UAVSAR regimes are calibrated from real interferograms in terms of coherence, posting, phase noise, and, for NISAR, the large-scale ionospheric spectrum. The benchmark also provides per-edge five-arc ambiguity labels, difficulty strata, and whole-tile DEM splits that prevent terrain leakage. Simulated residue densities are close to or moderately higher than those measured in the calibration data. Across eleven classical, network-flow, and deep-learning baselines, performance separates strongly by sensor regime and difficulty. A model trained only on LB-DLPU transfers without adaptation to real NISAR and UAVSAR interferograms; on a San Andreas UAVSAR scene, its output agrees with the JPL operational product to within one fringe over 95.9% of the scene. LB-DLPU, its frozen splits, generation code, and evaluation harness are publicly released to support reproducible assessment of L-band deformation-retrieval reliability.
- 11 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 23 (1545598X) , 1-5
A Method for Estimating Winter Wheat Height Using UAV Point Cloud Data Enhanced by Density Consistency Filtering
Point clouds generated by Structure from Motion (SfM) are often affected by significant noise caused by plant movement, as well as missing ground points caused by canopy occlusion. This reduces the quality of canopy and terrain extraction and makes it challenging to accurately estimate the height of crops from Unmanned Aerial Vehicle (UAV) images. To overcome these limitations, this study introduces a Density Consistency Filtering (DCF) algorithm, which adaptively models local density continuity to distinguish between valid points and noisy points. It effectively preserves local canopy structures while removing clustered noise. Furthermore, a color-spatial interpolation scheme based on ExG-RANSAC is developed to reconstruct missing ground points under dense canopies. Evaluated on six datasets from May to June 2019 covering key stages of winter wheat growth, the method achieved an RMSE of 7.7 cm, MAE of 6.2 cm, and R² of 0.91. After the early stem elongation stage, the method achieved an RMSE of 4.9 cm. The approach significantly improves estimation accuracy of the late growth stages, demonstrating strong potential for precision agriculture applications.
- 11 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 23 (1545598X) , 1-5
A Method for Estimating Winter Wheat Height Using UAV Point Cloud Data Enhanced by Density Consistency Filtering
Point clouds generated by Structure from Motion (SfM) are often affected by significant noise caused by plant movement, as well as missing ground points caused by canopy occlusion. This reduces the quality of canopy and terrain extraction and makes it challenging to accurately estimate the height of crops from Unmanned Aerial Vehicle (UAV) images. To overcome these limitations, this study introduces a Density Consistency Filtering (DCF) algorithm, which adaptively models local density continuity to distinguish between valid points and noisy points. It effectively preserves local canopy structures while removing clustered noise. Furthermore, a color-spatial interpolation scheme based on ExG-RANSAC is developed to reconstruct missing ground points under dense canopies. Evaluated on six datasets from May to June 2019 covering key stages of winter wheat growth, the method achieved an RMSE of 7.7 cm, MAE of 6.2 cm, and R² of 0.91. After the early stem elongation stage, the method achieved an RMSE of 4.9 cm. The approach significantly improves estimation accuracy of the late growth stages, demonstrating strong potential for precision agriculture applications.
- 1 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1430-1433
A Revised ICA Algorithm to Remove Cirrus Cloud Effects in Spectral Data of Landsat-8 Bands 1–7
A revised independent component analysis (ICA) algorithm was developed to improve the performance in the removal of cirrus cloud effects within spectral data of Landsat-8 Bands 1–7. The prior was that the reflectance values in the cirrus spectral band should be zero or near zero after the cloud removal. Then, the cloud component derived using the forward ICA transformation was revised and use to remove the cloud effects. After the inverse ICA transformation, the cloud removal was completed. Applying the algorithm to the Landsat-8 sub-image of 041/036 (path/row) acquired on 14 March 2015, we noted that cirrus clouds in Bands 1–7 disappeared visually. The cloud-free sub-image of the same location acquired on 1 April was extracted and used as the reference image in the algorithm verification. After the algorithm, R 2 values of the March image and reference image increased significantly. The revised algorithm was effective and valid. The revised one outperformed the original ICA algorithm in the comparison of both algorithms.
- 1 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1430-1433
A Revised ICA Algorithm to Remove Cirrus Cloud Effects in Spectral Data of Landsat-8 Bands 1–7
A revised independent component analysis (ICA) algorithm was developed to improve the performance in the removal of cirrus cloud effects within spectral data of Landsat-8 Bands 1–7. The prior was that the reflectance values in the cirrus spectral band should be zero or near zero after the cloud removal. Then, the cloud component derived using the forward ICA transformation was revised and use to remove the cloud effects. After the inverse ICA transformation, the cloud removal was completed. Applying the algorithm to the Landsat-8 sub-image of 041/036 (path/row) acquired on 14 March 2015, we noted that cirrus clouds in Bands 1–7 disappeared visually. The cloud-free sub-image of the same location acquired on 1 April was extracted and used as the reference image in the algorithm verification. After the algorithm, R 2 values of the March image and reference image increased significantly. The revised algorithm was effective and valid. The revised one outperformed the original ICA algorithm in the comparison of both algorithms.
- 1 July 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 915-918
Using Independent Component Analysis and Estimated Thin-Cloud Reflectance to Remove Cloud Effect on Landsat-8 Oli Band Data
Coupled with the independent component analysis (ICA) and estimated thin-cloud reflectance data, an algorithm to remove thin-cloud effects on the operational land imager (OLI) data of Landsat-8 was studied. The algorithm was evaluated with a subimage extracted from Landsat-8 041/036 scene acquired on 16 March 2015. Thin clouds were effectively removed visually after the algorithm. In the algorithm validation, a cloud-free Landsat-8 image acquired on 1 April 2015 was selected as the “truth”. Spatial correlation coefficients of the March image before the cloud-removal algorithm and April image and of the March image after the algorithm and April image were evaluated on the band-by-band basis. The coefficients increased significantly after the algorithm. Thus, the developed algorithm was quantitatively effective and valid.
- 1 July 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 915-918
Using Independent Component Analysis and Estimated Thin-Cloud Reflectance to Remove Cloud Effect on Landsat-8 Oli Band Data
Coupled with the independent component analysis (ICA) and estimated thin-cloud reflectance data, an algorithm to remove thin-cloud effects on the operational land imager (OLI) data of Landsat-8 was studied. The algorithm was evaluated with a subimage extracted from Landsat-8 041/036 scene acquired on 16 March 2015. Thin clouds were effectively removed visually after the algorithm. In the algorithm validation, a cloud-free Landsat-8 image acquired on 1 April 2015 was selected as the “truth”. Spatial correlation coefficients of the March image before the cloud-removal algorithm and April image and of the March image after the algorithm and April image were evaluated on the band-by-band basis. The coefficients increased significantly after the algorithm. Thus, the developed algorithm was quantitatively effective and valid.
- 1 July 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 927-930
Extendibility of a Thin-Cloud Removal Algorithm to Hi-Resolution Visible Bands of Sentinel-2 Data
The RTM-based algorithm developed to remove thin-cloud effects for Landsat-8 visible bands was extended to those of Sentine-2A visible bands. In the assessment of the extendibility, a cloud-covered Sentinel-2A image acquired on 18 February 2017 was downloaded. Another cloud-free Sentinel-2A image of the same area acquired on 8 February 2017 was selected as reference image. The three assumptions made in the original RTM-based algorithm were evaluated, and were valid for Sentinel-2A data. The algorithm was qualitatively and quantitatively effective in the thin-cloud removal for Sentinel-2A data of visible bands. Therefore, the algorithm developed for Landsat sensors was directly applicable to Sentinel-2 spectral data.
- 1 July 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 927-930
Extendibility of a Thin-Cloud Removal Algorithm to Hi-Resolution Visible Bands of Sentinel-2 Data
The RTM-based algorithm developed to remove thin-cloud effects for Landsat-8 visible bands was extended to those of Sentine-2A visible bands. In the assessment of the extendibility, a cloud-covered Sentinel-2A image acquired on 18 February 2017 was downloaded. Another cloud-free Sentinel-2A image of the same area acquired on 8 February 2017 was selected as reference image. The three assumptions made in the original RTM-based algorithm were evaluated, and were valid for Sentinel-2A data. The algorithm was qualitatively and quantitatively effective in the thin-cloud removal for Sentinel-2A data of visible bands. Therefore, the algorithm developed for Landsat sensors was directly applicable to Sentinel-2 spectral data.
- 1 July 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 4677-4680
Thin cloud detection using spectral similarity in coastal and blue bands of Landsat-8 data
With the assumption that spectral characteristics of coastal band (Band 1) and blue band (Band 2) of Landsat-8 were similar in a cloud-free area, an algorithm to detect thin clouds was derived. The assumption and algorithm was evaluated using Landsat-8 data of p14/r36 acquired on 1 April 2014. First, the assumed spectral similarity of Bands 1 and 2 was validated in the cloud-free area. The similarity was adversely affected once there was cloud (over the area). Second, the algorithm was applied to detect thin clouds in a full scene of Landsat-8 image. Visual comparison with the true color composite, the detected and observed clouds were very similar spatially. Next, the spatial correlation coefficient of the detected could image and the top-of-atmosphere (TOA) reflectance data of Band 9 of Landsat-8 was computed. The coefficient was 0.9604. Finally, with the least-squared-fit analysis, R2 value of 0.9223 was obtained. Since the interrupt of the least-squared-fit equation was -0.0003 or close to 0, there was the simple ratio relationship of the detected cloud and reflectance value of Band 9. Thus, the algorithm was not only able to detect thin clouds, but also to assess them quantitatively.
- 1 July 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 4677-4680
Thin cloud detection using spectral similarity in coastal and blue bands of Landsat-8 data
With the assumption that spectral characteristics of coastal band (Band 1) and blue band (Band 2) of Landsat-8 were similar in a cloud-free area, an algorithm to detect thin clouds was derived. The assumption and algorithm was evaluated using Landsat-8 data of p14/r36 acquired on 1 April 2014. First, the assumed spectral similarity of Bands 1 and 2 was validated in the cloud-free area. The similarity was adversely affected once there was cloud (over the area). Second, the algorithm was applied to detect thin clouds in a full scene of Landsat-8 image. Visual comparison with the true color composite, the detected and observed clouds were very similar spatially. Next, the spatial correlation coefficient of the detected could image and the top-of-atmosphere (TOA) reflectance data of Band 9 of Landsat-8 was computed. The coefficient was 0.9604. Finally, with the least-squared-fit analysis, R2 value of 0.9223 was obtained. Since the interrupt of the least-squared-fit equation was -0.0003 or close to 0, there was the simple ratio relationship of the detected cloud and reflectance value of Band 9. Thus, the algorithm was not only able to detect thin clouds, but also to assess them quantitatively.
- 1 July 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 48, 2381-2384
A new PolSAR decomposition algorithm to delineate urban targets
A new four component decomposition algorithm for urban targets delineation from PolSAR imagery was studied. First, a new correlation coefficient was introduced to describe characteristics of urban targets. The coefficient was the linear combination of two elements of a coherent matrix. Then, the coefficient was used to modify the volumetric scattering model component in Yamaguchi et al.'s four-component decomposition algorithm with the azimuth deorientation. Thus, a new four-component decomposition algorithm was established. With the algorithm, the percentage of double-bounced scattering increased but the percentage of volumetric scattering decreased in urban areas. Thus, the urban targets whose azimuth angles are 0° or non-zero° were effectively identified.
- 1 July 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 48, 2381-2384
A new PolSAR decomposition algorithm to delineate urban targets
A new four component decomposition algorithm for urban targets delineation from PolSAR imagery was studied. First, a new correlation coefficient was introduced to describe characteristics of urban targets. The coefficient was the linear combination of two elements of a coherent matrix. Then, the coefficient was used to modify the volumetric scattering model component in Yamaguchi et al.'s four-component decomposition algorithm with the azimuth deorientation. Thus, a new four-component decomposition algorithm was established. With the algorithm, the percentage of double-bounced scattering increased but the percentage of volumetric scattering decreased in urban areas. Thus, the urban targets whose azimuth angles are 0° or non-zero° were effectively identified.
- 3 November 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 570-573
Simulation of thin clouds in visible spectrum using the simplified radiative transfer equation
An algorithm to simulate the reflectance of clouds using visible bands was developed based on the simplified radiative transfer equation and two assumptions. The assumptions and algorithm were evaluated using a Landsat 8 sub-image of 015/035 (path/row) acquired on 27 August 2013. The relationship of visible bands in clear regions was verified to be linear. The effects of clouds in visible bands were linear as well. Then, the algorithm was used to simulate thin clouds. Band-by-band, the simulated clouds were in agreement with the distribution observed in each visible band. Finally, spatial correlation coefficients of simulated clouds and clouds detected by band 9 of Landsat-8 were calculated. High spatial coefficients quantitatively indicated the validity of the algorithm.
- 3 November 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 570-573
Simulation of thin clouds in visible spectrum using the simplified radiative transfer equation
An algorithm to simulate the reflectance of clouds using visible bands was developed based on the simplified radiative transfer equation and two assumptions. The assumptions and algorithm were evaluated using a Landsat 8 sub-image of 015/035 (path/row) acquired on 27 August 2013. The relationship of visible bands in clear regions was verified to be linear. The effects of clouds in visible bands were linear as well. Then, the algorithm was used to simulate thin clouds. Band-by-band, the simulated clouds were in agreement with the distribution observed in each visible band. Finally, spatial correlation coefficients of simulated clouds and clouds detected by band 9 of Landsat-8 were calculated. High spatial coefficients quantitatively indicated the validity of the algorithm.
- 1 June 2016
- Published by Elsevier in Remote Sensing of Environment
- Vol. 179, 183-195
An empirical and radiative transfer model based algorithm to remove thin clouds in visible bands
- 1 June 2016
- Published by Elsevier in Remote Sensing of Environment
- Vol. 179, 183-195
An empirical and radiative transfer model based algorithm to remove thin clouds in visible bands